In recent years, deep learning techniques have shown impressive performance in the field of identification of diseases of crops using digital images. In this work, a deep learning approach for identification of in-field diseased images of maize crop has been proposed. The images were captured from experimental fields of ICAR-IIMR, Ludhiana, India, targeted to three important diseases viz. Maydis Leaf Blight, Turcicum Leaf Blight and Banded Leaf and Sheath Blight in a non-destructive manner with varied backgrounds using digital cameras and smartphones. In order to solve the problem of class imbalance, artificial images were generated by rotation enhancement and brightness enhancement methods. In this study, three different architectures based on the framework of ‘Inception-v3’ network were trained with the collected diseased images of maize using baseline training approach. The best-performed model achieved an overall classification accuracy of 95.99% with average recall of 95.96% on the separate test dataset. Furthermore, we compared the performance of the best-performing model with some pre-trained state-of-the-art models and presented the comparative results in this manuscript. The results reported that best-performing model performed quite better than the pre-trained models. This demonstrates the applicability of baseline training approach of the proposed model for better feature extraction and learning. Overall performance analysis suggested that the best-performed model is efficient in recognizing diseases of maize from in-field images even with varied backgrounds.

As the largest radio telescope in the world, the Square Kilometre Array (SKA) will lead the next generation of radio astronomy. The feats of engineering required to construct the telescope array will be matched only by the techniques developed to exploit the rich scientific value of the data. To drive forward the development of efficient and accurate analysis methods, we are designing a series of data challenges that will provide the scientific community with high-quality datasets for testing and evaluating new techniques. In this paper we present a description and results from the first such Science Data Challenge (SDC1). Based on SKA MID continuum simulated observations and covering three frequencies (560 MHz, 1400MHz and 9200 MHz) at three depths (8 h, 100 h and 1000 h), SDC1 asked participants to apply source detection, characterization and classification methods to simulated data. The challenge opened in November 2018, with nine teams submitting results by the deadline of April 2019. In this work we analyse the results for 8 of those teams, showcasing the variety of approaches that can be successfully used to find, characterise and classify sources in a deep, crowded field. The results also demonstrate the importance of building domain knowledge and expertise on this kind of analysis to obtain the best performance. As high-resolution observations begin revealing the true complexity of the sky, one of the outstanding challenges emerging from this analysis is the ability to deal with highly resolved and complex sources as effectively as the unresolved source population.

We study the alignment of radio galaxies axes using the FIRST catalogue. we impose several cuts in order to select the candidates which are most likely to be free of systematic bias. In our study we primarily focus on testing for alignment among sources within a certain angular separation from one another since for most sources redshift information is not available. We find a very significant effect for angular distances less than 1 degrees. The distance scale of alignment is found to be roughly 28 Mpc, in agreement with earlier estimates, assuming that these sources are dominantly at redshift of 0.8. However, we are not able to entirely rule out the possibility of systematic bias in data. We also perform a full three dimensional analysis using a smaller data sample for which redshift information is available. In this case we only find a very weak signal at much larger distances.

The cosmological principle states that the Universe is statistically homogeneous and isotropic at large distance scales. There currently exist many observations which indicate a departure from this principle. It has been shown that many of these observations can be explained by invoking superhorizon cosmological perturbations and may be consistent with the Big Bang paradigm. Remarkably, these modes simultaneously explain the observed Hubble tension, i.e., the discrepancy between the direct and indirect measurements of the Hubble parameter. We propose several tests of the cosmological principle using SKA. In particular, we can reliably extract the signal of dipole anisotropy in the distribution of radio galaxies. The superhorizon perturbations also predict a significant redshift dependence of the dipole signal which can be nicely tested by the study of signals of reionization and the dark ages using SKA. We also propose to study the alignment of radio galaxy axes as well as their integrated polarization vectors over distance scales ranging from a few Mpc to Gpc. We discuss data analysis techniques that can reliably extract these signals from data.

Abstract:The world of technology is growing by leaps and bounds and the arena in technology that is going to be explored is Data Mining. It is estimated that till 2025, most of the world's trade will be based on Data Mining [1]. There is vast availability of people opinion data on twitter for almost every product and service. The challenge is to interpret this data and to extract the information which can lead a decision maker to take better decisions. In dictionary-based approach every word with some positive, negative or neutral value is mapped but opinions are not always direct, hence the sense of the sentence or sub-sentence doesn't agree with its numeral weight. This short coming of this approach lead us to come up with some strategies to increase the accuracy of this method by multiplying the weights together and using some fundamental semantic rule to classify sarcastic tweets. Hence in this paper a hybrid approach is implemented which ensures the sign of total weight of the sentence according to its indirect sense. The positive outcome is that opinions which were earlier treated as neutral are now retaining their sense and add up to our decisions. The hybrid approach is using the concepts of dictionary-based approach and semantic-based approach i.e. matching words from the dictionary and assigning their sentimental value and also using some specific semantic rules used for analyzing sarcastic or neutral tweets for gaining more information about the opinions. The proposed mining of opinions has become easier and more accurate that can be utilized for product's sale forecasting.